Abstract: This paper presents a lane recognition method which can satisfy the high-speed requirement of outdoor mobile robots.Firstly,we use edge extraction method and dynamic bi-threshold binarization algorithm to get a binary image from the original picture,then the ATN-based image comprehension method is used to get the position of lane.The method is used in THMR-V(TsingHua Mobile Robot-V) which also successfully achieves a maximal autonomous driving speed of 150km/h in highway.